The decrease in life insurance ownership: Implications
for financial planning

Kyoung Tae Kima, Travis P. Mountainb, Sherman D. Hannac,*,
Namhoon Kimd

aDepartment of Consumer Sciences, 312 Adams Hall, Box 870158, University of Alabama, Tuscaloosa,
AL 35487, USA

bDepartment of Agricultural and Applied Economics, Virginia Tech University, 250 Drillfield Drive,
Blacksburg, VA 24061, USA

cDepartment of Human Sciences, Ohio State University, 1787 Neil Avenue, Columbus, OH 43210, USA
dKorea Rural Economic Institute, 601 Bitgaram-ro, Naju-si. Jeollanam-do, 58321, South Korea

Abstract

Based on our analyses of Survey of Consumer Finances datasets, the proportion of households
owning a life insurance policy decreased from 72% in 1992 to 60% in 2016. We estimated logistic
regressions on the likelihood of ownership of any, term, and cash value life insurance. We conclude
that changes in household characteristics accounted for the decrease in term life insurance ownership,
but not for the decreases in any and in cash value life insurance ownership. We also found a positive
association between use of a financial planner and life insurance ownership. We discuss implications
for financial planning. © 2020 Academy of Financial Services. All rights reserved.

JEL classification: D12; D14; G22

Keywords: Life insurance; Declining insurance demand; Survey of Consumer Finances

1. Introduction

Life insurance is an important component of risk management and insurance planning,
and is also important in other financial planning topics, including employee benefits,
income tax planning, and estate planning. “As late as 1960, life insurance was the

* Corresponding author. Tel.: �1-614-292-4584; fax: �1-614-292-4339.
E-mail address: Hanna.1@osu.edu (S.D. Hanna)

Financial Services Review 28 (2020) 1-16

1057-0810/20/$ – see front matter © 2020 Academy of Financial Services. All rights reserved.



substance of financial planning. . . ” (Brandon and Welch, 2009, p. 2). However, the
importance of life insurance seems to be decreasing in terms of ownership rates, and the
relative number of life insurance agents (Scism, 2016). Also, the Certified Financial
Planner Board (2015) list of principal knowledge topics weight for risk management and
insurance for the CFP Exam, has decreased from 14% (Hanna, et al., 2011) to 12%
(Certified Financial Planner Board, 2015). A Wall Street Journal article (Scism, 2016)
suggested that because of consolidation and decreasing ownership rates, the life insur-
ance agent may be going the way of the dinosaur. Cordell, Finke, and Lemoine (2007)
noted the long-term trend of decreasing ownership of cash value life insurance between
1995 and 2004, and Retzloff (2010) reported that life insurance ownership in the United
States was at a 50 year low. However, life insurance is still a vital safety net for U.S.
households. The premature death of a wage-earner is one of the more serious financial
risks a household faces. Income replacement and burial expenses are the top two reasons
households own life insurance (Durham, 2015).

The main purpose of our research was to ascertain factors related to the decline of life
insurance ownership rates of U.S. households. We analyzed a combination of 1992 to
2016 Survey of Consumer Finances (SCF) datasets, and found that ownership of life
insurance decreased until 2013 and then stayed about the same in 2016. We conducted
logistic regression analyses of the pooled dataset to analyze the extent to which changes
in the composition of U.S. households might have contributed to the decreases in any,
term, and cash value life insurance. Further, we conducted additional analyses to
investigate the role of financial planner use on life insurance ownership. Our study
contributes to some implications for financial planners by providing insights into life
insurance ownership trends.

2. Literature review

2.1. Overview of life insurance

Without life insurance, most families would need to reduce their current standard of living
in the event of the death of a spouse or partner (Auerbach and Kotlikoff, 1991; Bernheim,
Carman, Gokhale, and Kotlikoff, 2003; Bernheim, Forni, Gokhale, and Kotlikoff, 2003).
Cash value and term life insurance are the two main categories of life insurance. In addition
to the insurance aspect of a cash value life policy, it also provides a savings component and
is intended to last the insured’s “whole” life. Term policies are pure insurance that are
associated with a set number of years ranging from one to 40 with 20 years being a common
term contract length. If the insured dies during this contract period, the beneficiary receives
the face amount of the policy. If one dies after the contract expires, no benefits are paid to
the beneficiary. Employer-provided group life insurance is one of the most common types of
employer benefits, and in 2017, 60% of employers provided it, with 73% of eligible
employees participating in group life insurance (Greenwald and Fronstin, 2019).

2 K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



2.2. Previous studies on life insurance ownership

Mulholland, Finke, and Huston (2016) examined the declining trend in cash value life
insurance ownership rates using the 1992 to 2010 waves of the SCF. The proportion of cash
value life policies relative to term life policies dramatically decreased over this period. They
found a substitution effect, as those who owned term life insurance were less likely to own
cash value life insurance. However, Frees and Sun (2010) found a complementary effect, as
those who owned term life insurance were more likely to own cash value life insurance.
Mulholland et al. (2016) and Glazer (2007) noted that in addition to the income protection
features of term life insurance, cash value life insurance policies have been marketed as
tax-advantaged investments. However, with the introduction of more tax-advantaged savings
instruments such as Roth IRAs, along with the introduction of tax-advantaged savings plans
for college costs, many households now have attractive alternatives to cash value life
insurance for important financial goals. Increases in federal income tax marginal rates could
potentially decrease the demand for cash value life insurance. Mulholland et al. (2016) noted
that cash value life insurance has had an important role in estate planning tools, especially
when more households were potentially subject to the federal estate tax. The increases in the
exemption amounts for the federal estate tax have generally reduced the number of house-
holds potentially subject to the tax after 2004. Mulholland et al. (2016) also noted that the
cost of term life insurance has dropped substantially since the introduction of internet
marketing and price comparisons, but the cost of cash value life insurance has not dropped
as much. Heo, Grable, and Chatterjee (2013) examined the 2004 and 2008 National Longi-
tudinal Survey of Youth 1979 cohort that consisted of respondents aged 43–51. Heo et al.
(2013) noted that there was a net decrease of one percentage in life insurance ownership over
this period.

Guillemette, Hussein, Phillips, and Martin (2015) analyzed the 1992 through 2010 SCF
datasets, finding a decreasing ownership trend and examined if household size affected life
insurance ownership differently for minority households. Compared with White households,
as household size increases, the likelihood of life insurance ownership decreases for both
Black and Hispanic households. Additionally, they found that being employed, age, educa-
tion, presence of children, homeownership, and being married to be positively associated
with having life insurance. Relative to 1992, all subsequent survey years were significant and
negative with the exception of 1995 that was not significantly different than 1992. They also
found that self-employed households were less likely to own life insurance compared with
those working for an employer, and Hispanic households were less likely to own life
insurance than otherwise comparable White households, while Black households were more
likely to own life insurance than White households.

Gutter and Hatcher (2008) examined the demand for life insurance using the 2004 SCF,
with an objective of determining factors related to White and Black household life insurance
ownership differences. Gutter and Hatcher (2008) found age to be positively correlated to life
insurance ownership. Homeownership was positively related to life insurance ownership
while the likelihood of life insurance ownership for those with children was not different
from those without children. Mountain (2015), using the 2013 SCF, examined life insurance
ownership for working coupled households aged 30–64. Mountain (2015) estimated that the

3K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



median proportion of insurable human wealth insured with life insurance was only 28%. Age
and education were positively associated with life insurance ownership. Households with a
Black respondent were more likely, and those with Hispanic and Asian/other respondents
were less likely, to own life insurance than households with a White respondent. Married
couples were more likely to own life insurance than partnered couples.

Liebenberg, Carson, and Dumm (2012) explored the determinants of life insurance
demand for both term and whole life policies in a dynamic analysis where the authors used
1983–1989 SCF panel datasets in a Cragg model. They found that new parents were more
likely to own life insurance than those who were not new parents, and households who had
started a new job were more likely to increase their holdings of term insurance coverage than
those who had not. Bernheim, Carman et al. (2003) examined financial vulnerability and life
insurance coverage at all stages of the life cycle. The authors proposed two explanations on
why a household with financial vulnerabilities might not be protected by life insurance: (1)
young households purchased long term life insurance contracts but failed to update them as
life events change; (2) actual needs had no effect on purchase.

Finke, Huston, and Waller (2009) used data from the 2004 SCF and compared the
difference of life insurance purchases between people who were advised by either financial
planners or brokers (dealers) to those who were not. They found that households that used
financial planners and that used broker/dealers shared similar demographic characteristics.
More important, the estimation results showed that people who relied primarily on financial
planners were more likely to purchase adequate life insurance holdings, while the use of
broker/dealers had no impact on levels of life insurance. Finally, households who were
wealthier, self-employed, and home-owning tended to purchase more adequate life insurance
holdings. Baek and DeVaney (2005) developed a model of term and whole life insurance
ownership and amount of coverage, and tested the effects of human capital, bequest motives,
and risk variables, as well as other household characteristics. They found that ownership of
term life insurance decreased with age, was higher for middle income households than for
low income households, and higher for homeowners than for renters, but education was not
significantly related to term life insurance ownership. They also found that ownership of cash
value life insurance was highest for those over the age of 65, and there was also a positive
relationship with age controlling for other characteristics including life expectancy. Cash
value life insurance ownership was also negatively related to having excellent health,
positively related to the level of liquid assets, and those in the highest income tax bracket
were more likely to own than those in the lowest bracket. Ownership was not significantly
related to attitudes about leaving a bequest.

Using Finke et al. (2009) as framework, Scott and Gilliam (2014) focused on baby boomer
life insurance adequacy before and after the 2008 financial crisis. The authors used 2004 and
2010 SCF datasets to represent before and after the 2008 financial crisis. They found that the
use of a financial planner was higher among household with adequate life insurance. In
addition, having a financial planner acted as a positive predictor of life insurance adequacy
in 2004, while it had no impact on life insurance adequacy in 2010, after the 2008 financial
crisis. Mountain (2015) found that households who consulted a financial planner or broker
were more likely to own life insurance and also, given ownership, to have higher face value

4 K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



amounts, however, he did not find the financial planner or broker variable to be significant
when exploring the proportion of insurable human wealth that was insured by life insurance.

2.3. Overview of past research on life insurance

We reviewed selected research on factors related to life insurance ownership and ade-
quacy. While previous researchers have found a number of household characteristics to be
related to life insurance ownership, none of the studies have had a focus on identification of
factors related to changes in life insurance ownership over many years. We estimated models
of life insurance ownership with a combined dataset of 1992 through 2016 SCF datasets;
thus, using more recent data than previous research, some of which included datasets only
through 2010. By estimating a time trend controlling for household characteristics, we
investigated the question of whether the decrease in life insurance ownership has been
primarily because of changes in household characteristics, or to some other factors, such as
the increasing availability of alternative tax advantaged investment options such as 401k
accounts (Mulholland et al., 2016).

3. Methods

3.1. Data and sample selection

We used a pooled dataset from nine waves of the Survey of Consumer Finances (SCF),
from 1992 to 2016. The Federal Reserve Board (FRB) has released the SCF cross-sectional
dataset triennially since 1983. Hanna, Kim, and Lindamood (2018) provide detailed infor-
mation about using SCF datasets. For more information about specific variables, see Board
of Governors of the Federal Reserve System (2014, 2017). The total sample size of the
pooled dataset is 44,634. There were 883 cases where life insurance ownership was missing
(shadow variables with values over 90, see Hanna et al., 2018), and those cases were
excluded from our analyses. The final analytic sample was 43,751.

3.2. Measurement of variables

3.2.1. Dependent variable
The SCF has life insurance ownership variables for both term life insurance and cash

value life insurance. For our empirical analysis as well as the life insurance ownership trend,
we use three separate binary dependent variables: whether the respondent or any family
member has, term, cash value, or any life insurance policy (i.e., term or cash value). The
actual questions are presented in Appendix 1.

3.2.2. Independent variables
Following the existing literature on life insurance ownership, the set of independent

variables included survey year, age of the household head, age squared, marital status
(married, single male, single female, and partnered), education of the household head (less

5K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



than high school, high school, some college, bachelor degree, and postbachelor degree),
race/ethnicity (White, Black, Hispanic, and Asian/others), presence of a child under 18,
current income compared with normal income (low, normal, and high), log of household
income, home ownership, employment status of the household head (salary worker, self-
employment, retired, and not working) and health status of a household head (excellent,
good, fair, and poor). In the SCF, race/ethnicity is of the respondent (Lindamood, Hanna, and
Bi, 2007), but for convenience, we refer to the race/ethnicity of the household, for example,
White households.

The SCF has had a question about whether a household reported using a financial planner
for information on (1) saving and investment decisions and/or (2) borrowing and credit
decisions since the 1998 SCF. We used an indicator for comprehensive financial planner use,
defined as a household using a financial planner for saving/investment and/or borrowing/
credit issues (Elmerick, Montalto, and Fox, 2002).

3.2.3. Analysis
Given the binary dependent variables of life insurance ownership, logistic regression

models are utilized to analyze factors related to the ownership of any life insurance, term life
insurance, and cash value life insurance. Because of the nature of the data and our analysis,
we are not able to attribute specific causal relationships for life insurance ownership. For
descriptive purposes, means tests are also employed. We used the Repeated-Imputation
Inference (RII) method with all of the five implicates in each SCF dataset, which provides
an estimate of variances more closely representing the true variances than estimates obtained
by only one implicate (Hanna et al., 2018; Lindamood et al., 2007). In order to provide
conservative hypothesis tests (Shin and Hanna, 2017), we did not weight the multivariate
analyses.

4. Results

4.1. Descriptive results

Fig. 1 displays the historical downward trend in life insurance ownership from the
combined 1992–2016 SCF datasets. The proportion of households owning some type of life
insurance peaked at 72% in 1992 and reached a low of 60% in 2013, with the 2016 rate about
the same as 2013. The proportion of households owning a term life insurance policy had a
generally downward trend, starting at 54% in 1992, and decreasing to 48% by 2013, and the
rate remained about the same in 2016. The proportion of households owning a cash value life
insurance policy decreased from 36% in 1992 to 20% in 2013, and the 2016 rate was about
the same as 2013. Table 1 presents the logistic regression results of the association between
survey year and the likelihood of three different types of life insurance ownership, not
controlling for other household characteristics. There were significant negative relationships
between the survey year and the likelihood of ownership of any life insurance, term life
insurance, and cash value life insurance.

6 K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



4.2. RII means test

The characteristics of our analytic sample are presented in Table 2. Most respondents were
White, homeowners, did not have a child under 18, had current income about the same as
normal, salary workers, completed education beyond the high school level and had good or
excellent health status. Results of means tests are also included to show descriptive patterns
of any, term, and cash value life insurance ownership by selected household characteristics.
In the discussion below, we report patterns for ownership of any life insurance. The life
insurance ownership rate was 46% for households with a respondent under 30, 71–72% for
those with a respondent 40 to 49 and 50 to 59, and 63% for those with a respondent 70 and
over. White households had a higher rate of life insurance ownership (69%) than each of the
other racial/ethnic groups, and Hispanics had the lowest rate, 37%. Homeowners had a higher
ownership (75%) than renters (46%).

The proportion of life insurance ownership was higher for married couples (77%) than for
other groups. Households with a dependent child under age 18 had a higher life insurance
ownership rate (68%) than those without a child (64%). Households with unusually high

Fig. 1. Trend in life insurance ownership rates, any type, term, and cash value, 1992–2016 SCF weighted results.
N � 43,751.

Table 1 Logistic regressions on ownership of any, term, and cash value life insurance by survey year only,
1992–2016 SCF

Any life insurance Term life insurance Cash value life insurance

Coefficient Standard
error

Coefficient Standard
error

Coefficient Standard
error

Survey year �0.0270*** 0.0030 �0.0072*** 0.0028 �0.0394*** 0.0030
Intercept 54.9622 6.0089 14.3913 5.5264 78.1899 6.0701
Mean concordance 50.1% 45.8% 53.1%

Note: *p � 0.05, **p � 0.01, ***p � 0.001. Unweighted analysis with RII technique. N � 43,751.

7K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



current income were more likely to own life insurance (71%) than those with lower current
income (54%) and those with current income about the same as normal (67%). Households
with the head working for a salary had a higher rate of life insurance ownership (71%) than
those in other types of employment status, and households with the head not working had
only a 34% life insurance ownership rate. The proportion of life insurance ownership was
highest for households with a head having a postbachelor degree (77%), compared with 75%
for households with a bachelor degree, 67% for households with some college, 63% for those
with a high school degree, and only 46% for those less than high school degree. Households
with a head with excellent health had the highest proportion of life insurance, 70%, while
those with poor health had the lowest proportion, 50%.

Table 2 Proportion of ownership of any, term, and cash value life, by household characteristics, 1992–2016
SCF

Variable Category Distribution
(%)

Any life insurance Term life insurance Cash value life insurance

Rate of
insurance
ownership
(%)

Significance
level

Rate of
insurance
ownership
(%)

Significance
level

Rate of
insurance
ownership
(%)

Significance
level

Age of head Less than 30 12.8 45.5 Reference 37.7 Reference 12.7 Reference
Between 30 and 39 19.0 63.7 �0.001 54.4 �0.001 20.0 �0.001
Between 40 and 49 20.5 71.1 �0.001 60.1 �0.001 23.6 �0.001
Between 50 and 59 18.0 71.5 �0.001 58.4 �0.001 28.0 �0.001
Between 60 and 69 14.0 70.2 �0.001 49.6 �0.001 33.5 �0.001
70 and over 15.7 62.9 �0.001 37.3 �0.001 32.7 �0.001

Race/ethnicity White 73.2 68.7 Reference 53.4 Reference 27.6 Reference
Black 13.5 66.5 �0.001 50.5 �0.001 24.7 �0.001
Hispanic 9.2 36.9 �0.001 31.8 �0.001 8.6 �0.001
Asian/other 4.1 59.0 �0.001 47.8 �0.001 18.8 �0.001

Homeowner No 33.7 45.5 Reference 36.6 Reference 13.7 Reference
Yes 66.3 75.0 �0.001 58.0 �0.001 30.9 �0.001

Marital status Married 50.3 77.1 Reference 61.4 Reference 31.6 Reference
Single male 14.9 51.7 �0.001 38.9 �0.001 18.2 �0.001
Single female 27.3 53.7 �0.001 39.9 �0.001 19.5 �0.001
Partner 7.5 52.2 �0.001 43.2 �0.001 16.2 �0.001

The presence
of child
under 18

No 66.2 63.8 Reference 47.4 Reference 26.1 Reference
Yes 33.8 67.6 �0.001 57.5 �0.001 23.3 �0.001

Current income
relative to
normal

Normal 73.1 67.1 Reference 52.3 Reference 26.1 Reference
Income higher 8.5 71.4 �0.001 56.9 �0.001 28.4 �0.001
Income lower 18.4 54.0 �0.001 42.2 �0.001 19.8 �0.001

Employment
status of
head

Salary worker 58.0 70.7 Reference 59.4 Reference 23.6 Reference
Self-employment 10.9 65.4 �0.001 48.8 �0.001 31.5 �0.001
Not working 5.7 33.8 �0.001 25.5 �0.001 12.5 �0.001
Retired 25.3 59.0 �0.001 37.6 �0.001 28.9 �0.001

Education of
household
head

less than high
school

14.6 45.6 Reference 33.5 Reference 16.3 Reference

High school degree 30.7 63.1 �0.001 47.7 �0.001 24.6 �0.001
Some college 23.0 66.6 �0.001 52.2 �0.001 25.4 �0.001
Bachelor degree 18.6 74.7 �0.001 61.0 �0.001 28.7 �0.001
Post-bachelor

degree
11.5 76.7 �0.001 61.6 �0.001 32.2 �0.001

Health status of
head

Excellent health 27.1 70.2 Reference 56.6 Reference 26.8 Reference
Good health 47.3 66.9 �0.001 52.9 �0.001 25.6 �0.001
Fair health 19.4 58.4 �0.001 42.8 �0.001 23.3 �0.001
Poor health 6.2 49.7 �0.001 34.9 �0.001 20.1 �0.001

Note: Author analyses of combined 1992–2016 SCF datasets, N � 43,751. RII technique is used for
significance tests. The reference category used in the means test is indicated in bold face. Significance test is for
mean difference from reference category for each variable.

8 K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



4.3. Multivariate results

Table 3 shows the logistic regression results of the likelihood of three different types of
life insurance ownership from the 1992–2016 SCF dataset, controlling for household char-
acteristics. There was a negative relationship between the survey year and the likelihood of
life insurance ownership for any life insurance and cash value life insurance, but the
relationship was not significant for term life insurance ownership. The magnitudes of the
coefficients of the time trend for ownership of any life insurance, and also for ownership of
cash value life insurance were very similar between the results in Table 3 and the results in
Table 1, suggesting that the negative trends in ownership and in cash value life insurance
were not because of changes in household characteristics. However, the effect of the time
trend on term life insurance ownership was not significant in Table 3, and the magnitude of

Table 3 Logistic regression analysis of likelihood of ownership of any, term, and cash value life,
1992–2016 SCF

Any life insurance Term life insurance Cash value life insurance

Coefficient Standard
error

Coefficient Standard
error

Coefficient Standard
error

Survey year �0.0277*** 0.0034 �0.0039 0.0030 �0.0422*** 0.0033
Age of head 0.0776*** 0.0093 0.0838*** 0.0092 0.0510*** 0.0108
Age squared (/10000) �6.3726*** 0.8658 �8.6443*** 0.8716 �2.6687** 0.9693
Race/ethnicity (reference: White)

Black 0.5309*** 0.0862 0.2519** 0.0793 0.3662*** 0.0903
Hispanic �0.9927*** 0.0972 �0.7450*** 0.0957 �0.8456*** 0.1414
Asian/other �0.3491** 0.1244 �0.2491* 0.1155 �0.2701* 0.1330

Homeowner 0.7555*** 0.0636 0.5241*** 0.0614 0.5487*** 0.0738
Marital status (reference: married)

Single male �0.7504*** 0.0793 �0.5418*** 0.0754 �0.4549*** 0.0864
Single female �0.7767*** 0.0689 �0.5186*** 0.0651 �0.5665*** 0.0753
Partnered �0.6230*** 0.1021 �0.4268*** 0.0970 �0.3280** 0.1205

The presence of child under 18 0.1236 0.0657 0.1713** 0.0582 0.0739 0.0646
Current income relative to normal

(reference: Normal)
Income Higher 0.0294 0.0898 �0.0272 0.0762 0.1150 0.0799
Income lower �0.3024*** 0.0671 �0.2439*** 0.0628 �0.1013 0.0724

Log of household income 0.0547*** 0.0138 0.0218 0.0129 0.0451** 0.0147
Employment status of head

(reference: Salary worker)
Self-employment �0.5130*** 0.0736 �0.6081*** 0.0624 0.2854*** 0.0646
Not working �1.1669*** 0.1182 �1.0809*** 0.1190 �0.3563* 0.1546
Retired �0.7338*** 0.0883 �0.6588*** 0.0793 �0.0608 0.0854

Education of head (reference: Less
than high school)

High school degree 0.3858*** 0.0842 0.1921* 0.0828 0.4210*** 0.1008
Some college 0.5439*** 0.0901 0.3173*** 0.0870 0.4984*** 0.1050
Bachelor degree 0.7273*** 0.0951 0.4433*** 0.0895 0.6248*** 0.1055
Post-bachelor degree 0.5862*** 0.1015 0.4375*** 0.0942 0.5248*** 0.1092

Health status of head (reference:
Excellent health)

Good health 0.0215 0.0615 0.0278 0.0537 0.0007 0.0574
Fair health �0.0697 0.0813 �0.0677 0.0747 �0.0392 0.0831
Poor health �0.2140 0.1256 �0.0945 0.1239 �0.2482 0.1422

Intercept 53.3805 6.7225 5.6963 6.0027 80.5698 6.5469
Mean concordance 75.6% 70.7% 72.0%

Note: *p � 0.05, **p � 0.01, ***p � 0.001. Unweighted analysis with RII technique. N � 43,751.

9K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



the effect was much less than the magnitude in Table 1, suggesting that the decrease in term
life insurance might have been because of changes in household characteristics from 1992 to
2016.

Both age and age squared were significantly related to the likelihood of insurance
ownership in each regression. The combined effect of the age variables implies that at the
2016 mean values of other independent variables, the likelihood of owning any life insur-
ance increased with age until age 61, then decreased with age; ownership of term life
insurance increased with age until age 48, then decreased with age; and ownership of cash
value life insurance increased with age for all ages in the sample. Fig. 2 shows the calculated
relationship between age of the head and the likelihood of ownership of each type of life
insurance, based on the coefficients in Table 3 and assuming the 2016 mean values of other
independent variables. This pattern is plausibly related to the cost of life insurance, which
increases substantially with age. Households who purchased 20-year term policies in their
40s and 50s would have expiring policies in their 60s and 70s. At this point in life the cost
of the insurance may outweigh the benefits, particularly if the household has relatively little
debt and has adequate assets to offset the loss of income resulting from the death of a spouse
or significant other. On the other hand, given estate planning goals and reduced opportunities
for tax sheltered retirement accounts after age 70, the estimated increase in ownership with
age of cash value ownership is plausible, since household income is assumed to remain at the
overall 2016 sample mean.

For the remaining independent variables in Table 3, we discuss only the effects for
ownership of any life insurance, though for many of the variables, the effects were similar
for ownership of term life insurance and of cash value life insurance. For racial/ethnic
patterns based on the regressions in Table 3, assuming other household characteristics had
the mean levels for 2016, Black households had the highest predicted likelihood of owning
life insurance, over 73%, compared with 61% for White households, 53% for Asian/other
households, and only 36% for Hispanic households. There is no theoretical reason for there

Fig. 2. Effect of age on the proportion of households owning any, term, and cash value life insurance, assuming
other household characteristics have 2016 mean levels. Estimates based on logistic regressions in Table 3.

10 K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



to be differences in life insurance ownership between difference race/ethnic groups, if
other household characteristics are equal. The reason for these differences may be related
to differences in marketing practices of insurance companies. Hispanic and Asian
households consist of a higher proportion of immigrant households, so it is possible that
insurance companies have not effectively found a way to market to these individuals
(Mountain, 2015), especially if life insurance is a product that is not well established in
their country of origin. Hispanic and Asian households may also rely more heavily on
their extended family network than both Black and White households, something that is
not measured in the SCF.

For the relationship between marital status and life insurance ownership, assuming
other household characteristics had the mean levels for 2016, married households had the
highest predicted likelihood of owning life insurance, almost 70%, compared with 56%
for partner households, 50% for single male households, and 49% for single female
households. Surviving individuals in a coupled household are likely to be financially
burdened if a spouse or partner were to die. Life insurance can help fill this burden. In
general, single headed households are not financially dependent on another wage-earner
and have little need for life insurance, because there would be no surviving spouse,
though as Nam and Hanna (2019) discussed, there might still be a need in terms of
making sure the person with custody of dependent children would have adequate
resources. The difference between married and partnered coupled households may be
because of the married households having a longer-term time horizon than otherwise
similar partnered households. Presence of a dependent child under age 18 was not
significantly related to ownership of any life insurance or of cash value life insurance,
but was positively related to ownership of term life insurance. This is reasonable,
because dependent children make the need for support for them paramount.

Household income was positively associated with the likelihood of owning any life or
of owning cash value life insurance. For the relationship between income shocks and life
insurance ownership, assuming other household characteristics had the mean levels for
2016, households with current income below normal had the lowest predicted likelihood
of owning life insurance, 54%, compared with 64% for households with higher than
normal income and 61% for households with normal income. This may be because of
household budget constraints. Households may accurately perceive the likelihood of
death in the current year as very low, and, thus, temporarily terminate a life insurance
policy if household income falls below normal income. Conversely, when household
income is above normal income, households may feel less of a constraint on the
household budget and use this extra income on life insurance, something they otherwise
were not willing to own. Current income was positively related to the likelihood of
owning any life insurance and to the likelihood of owning cash value life insurance, but
not to the likelihood of owning term life insurance. Given that employment status is
controlled, the lack of significance is reasonable, as an employer group life insurance is
something that is available for many employees without having to make a purchase
decision.

For the relationship between employment status of the head and life insurance ownership,
assuming other household characteristics had the mean levels for 2016, households headed

11K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



by an employee had the highest predicted likelihood of owning life insurance, 68%,
presumably partly because of the availability of group life insurance from many employers.
Group insurance may be provided for free or at reduced rate as an employer benefit.
Additionally, group life insurance removes the barrier of having to have a medical exam to
be insured, something that is common in the individual insurance market. This barrier may
prevent individuals from seeking life insurance, even if they would qualify for the policy
upon completion of the exam. The predicted life insurance ownership rates were only 51% for
households with a self-employed head, 49% for households of a retired head, and 40% for
households with heads not working but not of retirement age.

For the relationship between the education of the head and life insurance ownership,
assuming other household characteristics had the mean levels for 2016, households headed
by somebody with a postbachelor degree had a predicted life insurance ownership rate of
65%, and those with a bachelor degree had a predicted rate of 68%, compared with 62% for
those with some college but no degree, 57% for those whose highest education is a high
school degree, and 44% for those with no high school degree. Higher education levels are
likely related to more future oriented thinking (Yuh and Hanna, 2010), which would be
consistent with the life insurance ownership patterns shown in Fig. 10.

For the relationship between homeownership status and life insurance ownership, assum-
ing other household characteristics had the mean levels for 2016, homeowners had a
predicted ownership rate of 67%, compared with a predicted rate of 47% for renters.
Homeowners with a mortgage may need to have life insurance to make sure the loan is paid
off if a primary earner dies. Homeownership may also be a signal of financial stability.
Typically, homeowners are required to make a substantial down payment to purchase a
home, which, for most households requires substantial financial planning. Therefore, hom-
eowners may be more likely to both afford and plan for life insurance.

4.4. The association between use of financial planner and life insurance ownership

Table 4 shows the logistic regression results of the association between the use of
comprehensive financial planner and the likelihood of different types of life insurance

Table 4 Logistic regression analysis, the association between comprehensive financial planner use and the
likelihood of life insurance ownership, 1998–2016 SCF

Any life insurance Term life insurance Cash value life insurance

Coefficient Standard
error

Coefficient Standard
error

Coefficient Standard
error

Use of comprehensive
financial planner

0.3727*** 0.0425 0.2117*** 0.0356 0.2989*** 0.0364

Survey year �0.0230*** 0.0021 0.0169 0.0020 �0.0398*** 0.0022
Intercept 44.0071*** 4.2934 �5.6204 3.9183 75.8085*** 4.3208
Other control variables Yes Yes Yes
Mean concordance 75.9% 71.3% 71.7%

Note: *p � 0.05, **p � 0.01, ***p � 0.001. Unweighted analysis with RII technique. N � 35,667. Control
variables include the independent variables shown in Table 3.

12 K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



ownership, controlling for the other independent variables in Table 3. The SCF did not have
a specific code for use of financial planners before 1998, so we used a pooled dataset from
the 1998 –2016 SCF for this additional analysis. The utilization of a comprehensive
financial planner was positively associated with ownership of any, term, and cash value
life insurance policies. Households who used a comprehensive financial planner had
45.2% higher odds of owning any life insurance than those without using a financial
planner. Similarly, the use of comprehensive financial planner increased the odds of
owning term life and whole life insurance policy by 24% and 35%, respectively.
Controlling for use of a financial planner did not substantially change the effect of the
survey year compared with the results in Table 3, as the survey year had a negative effect
for any life insurance and for cash value life insurance, but the effect was not signifi-
cantly different from zero for term life insurance.

5. Discussion and implication

The effect of the time trend in the logistic regression means that even if the charac-
teristics of households in the United States had not changed between 1992 and 2016,
there would have been a substantial decrease in ownership of any life insurance and of
cash value life insurance during that period. However, based on the logistic regression
in Table 3, ownership of term life insurance might not have decreased if household
characteristics had not changed. The trends for some household characteristics related to
lower likelihood of life insurance ownership (Hispanic and Asian/other households,
unmarried households) might have contributed to lower ownership during the 1992 to
2016 period, but the trends for education and age might have contributed to increased
ownership. Overall, changes in household characteristics during the periods played a
relatively small role in the decreases in ownership of any and of cash value life
insurance, partly because most household characteristics changed slowly. For instance,
the proportion of households with married couples decreased from 53.9% in 1992 to
46.8% in 2016, and the proportion of household heads who were employees (vs.
self-employed or retired or not working) changed from 55% in 1992 to 56% in 2016.

Insuring human wealth is an appropriate life insurance objective (Chen, Ibbotson,
Milevsky, and Zhu, 2006) but even at the current ownership rates, the amount of life
insurance is not enough to replace lost human wealth for many households (Mountain,
2015). While the cash-value policy decline may reflect rational decision-making because
of changes in tax laws and the increase in tax advantaged investment options, these
patterns could change with changes in the federal income tax and with any substantial
changes in the federal estate tax.

Durham (2015) found that households often overestimate the cost of life insurance and
do so by drastic amounts. When households were asked to provide what they thought a
$250,000 20-year term policy would cost, one-quarter overestimated the cost by 625%
and one half overestimated the cost by 250%. Meanwhile, life insurance advice from a
financial planner is something that families place a low value on (Warschauer and
Sciglimpaglia, 2012), making life insurance planning something a family is more likely

13K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



to do on its own. While we found higher likelihood of life insurance ownership for
households who have a financial planner, we cannot be certain if the financial planner’s
advice led to life insurance ownership or if households who have life insurance were
more likely to seek advice from a financial planner. Future research might use methods
to correct for the selection effect, as was done by Kim, Pak, Shin, and Hanna (2018) for
analyses of the effect of financial planner use on holding a retirement savings goal.

It is of little surprise that life insurance ownership trends are decreasing when households
place little importance on life insurance planning while also dramatically overestimating its
cost. If financial planners were to view term life insurance as a loss leader product, or a
breakeven product, instead of a profit generator, both the financial planner and client might
be better off. As a gateway product, term insurance can help establish trust and rapport with
clients. Planners could start by asking the client how much they think a $100,000 20-year
term life policy would cost them, and see how that compares with reality. Once the client has
the peace of mind that his or family will be financially taken care of upon an untimely death,
relatively inexpensively, transition to other planning areas can ensue. This approach would
likely benefit financial planners who are following or not following a fiduciary standard,
recommendations of term life insurance without direct financial benefit may diminish
apprehension some clients might otherwise feel may have when presented with the advice to
purchasing life insurance. For clients with more complicated financial needs where cash-
value life insurance is appropriate, it may be a product that is brought up later, rather earlier,
in the financial plan.

If the ownership rates continue to decline, more and more families will face financial
hardship and economic distress if a spouse or partner dies prematurely. In terms of the
potential benefits of financial planning advice based on normative economic models, risk
management is an important component of the value of advice (Hanna and Lindamood,
2010), and for many families, life insurance still should be a salient component of a financial
plan. This is surprising as there are several studies have found that a substantial proportion
of households still remain underinsured. Financial planners and educators can better inform
their clients and the population as to the merits of life insurance. Additionally, alternative
products, adjustments to the current life insurance products, or additional saving will need to
fill the void of the disappearing life insurance.

Given our conclusion that the decreases in ownership of any life insurance and in
ownership of cash value life insurance were not because of changes in household charac-
teristics between 1992 and 2016, it is plausible that future changes in federal income tax and
estate tax policies could lead to future increases in life insurance ownership. For instance, if
a proposal by Senator Bernie Sanders to decrease the estate tax exemption from $11.4 million
in 2019 to $3.5 million (2009 level) were to be implemented, many wealthy households
would likely make substantial changes in their estate planning (Shenkman, 2019), and cash
value life insurance might provide one type of strategy. Further, proposed increases in federal
income tax rates might also make cash value life insurance more attractive. Even if no
near-term tax legislation is passed, most individual income tax changes in the Tax Cuts and
Jobs Act of 2017 are set to sunset January 1, 2026. The increases in income tax rates might
nudge households to reconsider cash-value life insurance policies.

14 K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



References

Auerbach, A. J., & Kotlikoff, L. (1991). Life Insurance Inadequacy-Evidence From a Sample of Older Widows
(NBER Working Papers 3765). National Bureau of Economic Research.

Baek, E., & DeVaney, S. A. (2005). Human capital, bequest motives, risk, and the purchase of life insurance.
Journal of Personal Finance, 4, 62–84.

Bernheim, B. D., Carman, K. G., Gokhale, J., & Kotlikoff, L. J. (2003). Are life insurance holdings related to
financial vulnerabilities? Economic Inquiry, 41, 531–554.

Bernheim, B. D., Forni, L., Gokhale, J., & Kotlikoff, L. J. (2003). The mismatch between life insurance holdings
and financial vulnerabilities: Evidence from the Health and Retirement Study. American Economic Review,
93, 354–365.

Board of Governors of the Federal Reserve System. (2014). Codebook for the 2013 Survey of Consumer
Finances. (available at http://www.federalreserve.gov/econresdata/scf/files/codebk2013.txt)

Board of Governors of the Federal Reserve System. (2017). Codebook for the 2016 Survey of Consumer
Finances. (available at https://www.federalreserve.gov/econres/files/codebk2016.txt)

Brandon, Jr., E. D., & Welch, H. O. (2009). The History of Financial Planning: The Transformation of Financial
Services. Hoboken, NJ: John Wiley & Sons.

Certified Financial Planner Board. (2015). 2015 Principal Knowledge Topics. (available at https://www.cfp.net/
become-a-cfp-professional/2015-job-task-analysis/2015-principal-knowledge-topics)

Chen, P., Ibbotson, R. G., Milevsky, M. A., & Zhu, K. X. (2006). Human capital, asset allocation, and life
insurance. Financial Analysts Journal, 62, 97–109.

Cordell, D. M., Finke, M., & Lemoine, C. (2007). The death of life? The demand for cash-value life insurance
has slipped in recent years, say the authors, much of it due to bad publicity. Journal of Financial Planning,
20, 42–44.

Durham, A. (2015). 2015 Insurance Barometer Study. Life Happens. (available at https://www.lifehappens.org/
industry-resources/2015barometer/)

Elmerick, S. A., Montalto, C. P., & Fox, J. J. (2002). Use of financial planners by U.S. households. Financial
Services Review, 11, 217–231.

Appendix 1 Life insurance variables in the SCF

Variable Description

X4001 Do you (or anyone in your family living here) have any life insurance? Please include
individual and group policies, but not accident insurance.

1. *YES
5. *NO

X4002 The two major types of life insurance are term and cash-value policies. Term policies pay a
benefit if the insured person dies, but otherwise have no value. They are often provided
through an employer or union, but may also be bought by individuals. Cash-value policies
also pay a death benefit, but differ in that they build up a value as premiums are paid. Are
any of your (family’s) policies term insurance?

1. *YES
5. *NO
0. Inap. (no life insurance: X4001∧ � 1)

X4004 Do you have any policies that build up a cash value or that you can borrow on? These are
sometimes called ‘whole life’, “straight life,” or “universal life” policies.

1. *YES
5. *NO
0. Inap. (no life insurance: X4001∧ � 1)

15K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16



Federal Reserve Bank. (2016). Codebook for 2016 Survey of Consumer Finances. Washington, DC: Federal
Reserve Bank.

Finke, M. S., Huston, S. J., & Waller, W. (2009). Do contracts impact comprehensive financial advice? (available
at http://dx.doi.org/10.2139/ssrn.1429807)

Frees, E. W., & Sun, Y. (2010). Household life insurance demand: A multivariate two-part model. North
American Actuarial Journal, 14, 338–354.

Glazer, D. A. (2007). The effect of commission versus non-commission benefits on customer value: The case of
life insurance policy performance. Financial Services Review, 16, 135–153.

Greenwald, L., & Fronstin, P. (2019). The state of employee benefits: Findings from the 2018 Health and
Workplace Benefits Survey. Employee Benefit Research Institute No. 470. (available at https://www.ebri.org/
docs/default-source/ebri-issue-brief/ebri_ib_470_wbs2-10jan19.pdf?sfvrsn�c5db3e2f_10)

Guillemette, M., Hussein, M., Phillips, G. M., & Martin, T. (2014). Minority household size and the life insurance
purchase decision. Financial Services Review, 24, 37–51.

Gutter, M. S., & Hatcher, C. B. (2008). Racial differences in the demand for life insurance. Journal of Risk and
Insurance, 75, 677–689.

Hanna, S. D., & Lindamood, S. (2010). Quantifying the economic benefits of personal financial planning.
Financial Services Review, 19, 111–127.

Hanna, S. D., Ji, H., Lee, J., Son, J., Letkiewicz, J., Lim, H., & Zhang, L. (2011). Content analysis of Financial
Services Review. Financial Services Review, 20 (3), 237–251.

Hanna, S. D., Kim, K. T., & Lindamood, S. (2018). Behind the numbers: Understanding the Survey of Consumer
Finances. Journal of Financial Counseling and Planning, 29, 410–418.

Heo, W., Grable, J. E., & Chatterjee, S. (2013). Life insurance consumption as a function of wealth change.
Financial Services Review, 22, 389–404.

Kim, K. T., Pak, T. K., Shin, S. H., & Hanna, S. D. (2018). The relationship between financial planner use and
holding a retirement saving goal: A propensity score matching analysis. Financial Planning Review, 1, e1008.
https://doi.org/10.1002/cfp2.1008

Liebenberg, A. P., Carson, J. M., & Dumm, R. E. (2012). A dynamic analysis of the demand for life insurance.
Journal of Risk and Insurance, 79, 619–644.

Lindamood, S., Hanna, S. D., & Bi, L. (2007). Using the Survey of Consumer Finances: Some methodological
considerations and issues. Journal of Consumer Affairs, 41, 195–222.

Mountain, T. P. (2015). Life Insurance and Financial Vulnerability. Ph.D. dissertation, The Ohio State
University.

Mulholland, B., Finke, M., & Huston, S. (2016). Understanding the shift in demand for cash value life insurance.
Risk Management and Insurance Review, 19, 7–36.

Nam, Y., & Hanna, S. D. (2019). The effects of risk aversion on life insurance ownership of single-parent
households. Applied Economics Letters, 26, 1285–1288. http://dx.doi.org/10.1080/13504851.2018.1546044

Retzloff, C. D. (2010). Household Trends in U.S. Life Insurance Ownership. LIMRA, report, November, 1.
Scism, L. (2016, Feb 25). Death of the insurance salesman at MetLife. The Wall Street Journal.
Scott, J. K., & Gilliam, J. (2014). Boomers’ life insurance adequacy pre & post the 2008 financial crisis. Financial

Services Review, 23, 287–304.
Shenkman, M. (2019, March 21). Sanders Estate Tax Proposal: Estate Planning Steps to Take Now.

(available at https://www.forbes.com/sites/martinshenkman/2019/03/21/sanders-estate-tax-proposal-es-
tate-planning-steps-to-take-now/#16bae6299f5e)

Shin, S. H., & Hanna, S. D. (2017). Accounting for complex sample designs in analyses of the Survey of
Consumer Finances. Journal of Consumer Affairs, 51, 255–283.

Warschauer, T. A. (2012). The economic benefits of personal financial planning: An empirical analysis. Financial
Services Review, 21, 195–208.

Yuh, Y., & Hanna, S. D. (2010). Which households think they save? Journal of Consumer Affairs, 44, 70–97.

16 K. Tae Kim et al. / Financial Services Review 28 (2020) 1-16